DOA Estimation Of Coherent Sources For MIMO Radar Based On Covariance Matrix Sparse Representation

Compared with the traditional phased radar,MIMO radar has the significantly improved array performance in terms of target recognition,detection and parameter estimation.DOA estimation is an important aspect of MIMO radar parameter estimation.The method for DOA estimation using covariance matrix sparse representation in MIMO radar can obtain lower computational complexity and higher angular resolution.However,when the sources are coherent,the covariance matrix of the sources is no longer a diagonal matrix.Therefore,the sparse representation model of covariance matrix based on Khatri-Rao product cannot be established.And then the sparse reconstruction method based on single measurement vector is unable to effectively achieve DOA estimation of coherent sources in MIMO radar.Aiming at solving the above-mentioned problems,the main researches are as follows:(1)To estimate DOA of coherent sources in MIMO radar with nonoverlapping virtual elements,a method based on covariance matching SL0 algorithm for DOA estimation of coherent sources in MIMO radar is proposed.A full rank covariance matrix with the property of Toeplitz is reconstructed using the covariance matching criteria.The high-order power of the inverse covariance matrix,which approximates the noise subspace,is used for calculating the weighted vector.Finally,the SL0 algorithm with lower computational complexity and higher reconstruction precision is used to estimate DOA by solving the model of covariance matrix sparse representation.Experimental results demonstrate that the proposed algorithm can effectively achieve DOA estimation of coherent signals without the number of sources.(2)When the virtual elements formed in MIMO radar are partially overlapped,the covariance matrix of virtual array in MIMO radar is often a block Toeplitz matrix.Therefore,the spatial smoothing method cannot be directly applied to solve the problem of coherent sources DOA estimation in MIMO radar.A method based on the block spatial smoothing and the improved SL0 algorithm for coherent sources DOA estimation in MIMO radar is proposed.Firstly,a group of Toeplitz sub-matrices can be obtained by dividing he covariance matrix according to the number of receive antennas.Then the spatial smoothing is used to recover the rank of each sub-matrix and solve the coherent problem between sub-matrices.Finally,the improved SL0 algorithm can be used to estimate DOA by solving the model of covariance matrix sparse representation.The method can effectively estimate DOA of coherent sources in MIMO radar with arbitrary interelement distances of transmit and receive antennas.(3)The software of coherent sources DOA estimation in MIMO radar based on LabVIEW is developed.Firstly,the parameters required for DOA estimation are realized through the input controls in LabVIEW.By using the MATLAB Script node in LabVIEW,the covariance matrix of virtual array in MIMO is preprocessed for handling coherent targets and the improved SL0 algorithm can be used to solve the problem of sparse signal reconstruction.Finally,the result is displayed by using the XY graph of LabVIEW.